Extended Endocrine Therapy and Breast Cancer Subtype

Key Takeaways

• Lower estimated invasive and distant recurrence risk was observed with extended endocrine therapy overall.
• Lower estimated risk was observed across luminal A-like, luminal B-like/ERBB2-negative, and ERBB2-positive disease, with the strongest distant recurrence signal in luminal A-like disease.
• Subgroup comparisons were limited by small numbers, overlapping confidence intervals, and the need for confirmation in larger cohorts.

Among 487 premenopausal patients with node-positive, hormone receptor-positive early breast cancer, extended endocrine therapy was associated with a lower estimated risk of distant recurrence than no extension in a JAMA Network Open cohort study. All patients had completed 5 years of LHRH agonist-based adjuvant endocrine therapy without recurrence, and the cause-specific HR for distant recurrence-free survival was 0.47 (95% CI, 0.30-0.75). Lower estimated risk was seen across surrogate subtypes, with the largest distant recurrence signal in luminal A-like disease.

The analysis combined two prospectively maintained datasets from the United States and Italy and included patients diagnosed at age 40 years or younger between January 2005 and December 2016. Eligible patients had node-positive, hormone receptor-positive, nonmetastatic early breast cancer, remained premenopausal after 5 years of therapy, and had no recurrence before the year-6 landmark. Exposure was defined on the first day of year 6 as tamoxifen monotherapy, LHRH agonist plus tamoxifen, or LHRH agonist plus aromatase inhibitor, with 276 patients exposed and 211 unexposed. Median age at diagnosis was 37 years in both groups, median follow-up was 7.3 years, and median extended therapy duration was 3.7 years. Subtype distribution was 18% luminal A-like, 61% luminal B-like/ERBB2-negative, and 21% ERBB2-positive, and propensity score weighting addressed baseline differences.

For invasive breast cancer-free survival, the PS-weighted HR was 0.64 (95% CI, 0.44-0.93) for extended therapy versus no extended therapy. The 5-year estimate was 85% with extended therapy and 79% without, with 52 and 71 invasive events during follow-up. First events included ipsilateral, contralateral, and local-regional recurrences, as well as distant recurrence and death. Across subtypes, the 5-year estimates were 78% versus 72% for luminal A-like disease, 84% versus 77% for luminal B-like/ERBB2-negative disease, and 97% versus 91% for ERBB2-positive disease. The direction of association was similar across all three surrogate subtypes.

At 5 years, the PS-weighted cumulative incidence of distant recurrence was 8% with extended therapy and 16% without. A total of 27 and 43 distant recurrence events occurred as the first event in the extended and unextended groups. The subtype-specific estimates were 6% versus 23% in luminal A-like disease, 10% versus 18% in luminal B-like/ERBB2-negative disease, and 3% versus 5% in ERBB2-positive disease. Cause-specific hazard ratios for distant recurrence-free survival were 0.25 (95% CI, 0.08-0.75), 0.54 (95% CI, 0.32-0.94), and 0.54 (95% CI, 0.12-2.53), respectively. Luminal A-like disease showed the greatest estimated distant recurrence benefit in this analysis.

The authors characterized the analysis as secondary and hypothesis-generating rather than definitive for subtype comparisons. Small subgroup counts, limited power, and 95% CIs that mostly included the point estimates of the other subgroups constrained interpretation, and subtype assignment relied on immunohistochemistry with uncertainty around Ki-67 and PgR cutoffs. Race and ethnicity data were unavailable.

The investigators concluded that the subtype pattern, especially the larger luminal A-like signal, should be confirmed in larger prospective cohorts.